Stingless Bee, Trigona Spinipes James C

Received 13 November 2003 Accepted 11 February 2004 Published online 15 June 2004

Olfactory eavesdropping by a competitively foraging stingless bee, Trigona spinipes James C. Nieh1*, Lillian S. Barreto2, Felipe A. L. Contrera3 and Vera L. Imperatriz-Fonseca3 1University of California San Diego, Division of Biological Sciences, Section of Ecology, Behavior and Evolution, Mail Code 0116, La Jolla, CA 92093, USA 2Empresa Baiana de Desenvolvimento Agricola, Laboratorio de Abelhas (LABE), Avenue Ademar de Barros N 967, Salvador-Bahia, Brazil 3Instituto de Biocieˆncias, Laborato´rio de Abelhas, Universidade de Sa˜o Paulo, Sa˜o Paulo 05508-900, SP, Brazil Signals that are perceived over long distances or leave extended spatial traces are subject to eavesdropping. Eavesdropping has therefore acted as a selective pressure in the evolution of diverse animal communication systems, perhaps even in the evolution of functionally referential communication. Early work suggested that some species of stingless bees (Hymenoptera, Apidae, Meliponini) may use interceptive olfactory eavesdropping to discover food sources being exploited by competitors, but it is not clear if any stingless bee can be attracted to the odour marks deposited by an interspecific competitor. We show that foragers of the aggressive meliponine bee, Trigona spinipes, can detect and orient towards odour marks deposited by a competitor, Melipona rufiventris, and then rapidly take over the food source, driving away or killing their competitors. When searching for food sources at new locations that they are not already exploiting, T. spinipes foragers strongly prefer M. rufiventris odour marks to odour marks deposited by their own nest- mates, whereas they prefer nest-mate odour marks over M. rufiventris odour marks at a location already occupied by T. spinipes nest-mates. Melipona rufiventris foragers flee from T. spinipes odour marks. This olfactory eavesdropping may have played a role in the evolution of potentially cryptic communication mechanisms such as shortened odour trails, point-source only odour marking and functionally referential communication concealed at the nest. Keywords: eavesdropping; olfaction; competition; aggression; evolution of communication

1. INTRODUCTION (Nieh 1999). Several species of stingless bees use extended odour trails to communicate food location (Kerr 1960). Eavesdropping plays a significant role in the evolution of Recently, an intermediate strategy was described for a complex and sophisticated communication networks in stingless bee, Trigona hyalinata, which uses a short odour animals ranging from songbirds to fishes (McGregor trail extending only a short distance from the food source 1993; Stowe et al. 1995; Oliveira et al. 1998; Peake et al. towards the nest (Nieh et al. 2003a). Other meliponine 2002; Whitfield 2002) and may play a role in shaping bee species, including those that may use functionally referen- communication (Nieh 1999). Signals that are sent over tial communication, do not use odour trails and odour long distances or leave extended spatial traces are suscep- mark the food source alone (Nieh & Roubik 1995; Hrncir tible to eavesdropping (McGregor 1993), and thus some et al. 2000; Nieh et al. 2003c). Intriguingly, honeybees extirpating stingless bee species (Hymenoptera, Apidae, odour mark only the food source and use functionally ref- Meliponini) are hypothesized to use interceptive olfactory erential communication, encoding food location through eavesdropping (intercepting signals intended for other a waggle dance at the nest (Esch et al. 2001; Dyer 2002). receivers (Peake 2004)) to detect valuable food sources Stingless bees are all highly social (Michener 2000), discovered by other bees (Kerr et al. 1963), whereas other occupy environments where food resources are seasonally species may detect the pheromones of aggressive melipon- scarce and can be highly sought after, and recruit nest- ine species to avoid costly conflicts with superior competi- mates to these resources (Johnson 1974; Roubik 1982; tors (Johnson 1974; Hubbell & Johnson 1978). Eltz et al. 2001, 2002; Liow et al. 2001). To help guide Such olfactory eavesdropping could have contributed to nest-mates during recruitment, certain species deposit the evolution of concealed communication inside the nest odour trails beginning near the nest and extending to the (Nieh 1999), and thus to the evolution of functionally ref- food source (Lindauer & Kerr 1958). Kerr (1960) reports erential communication (the ability to abstractly encode a T. amalthea odour trail extending for 900 m, and the environmental information into signals understood by foraging ranges of many meliponine species are thought receivers (Marler et al. 1992; Blumstein 1999)) as referen- to extend for at least several hundred metres (Roubik & tial location information replaced odour trail information Aluja 1983; Van Nieuwstadt & Ruano 1996). Such odour trails would create a long but relatively narrow active space and could be detected by scout bees whose search paths intersected the active space, with the probability of inter- * Author for correspondence ([email protected]). section increasing with trail length. The cross-sectional

Proc. R. Soc. Lond. B (2004) 271, 1633–1640 1633  2004 The Royal Society DOI 10.1098/rspb.2004.2717 1634 J. C. Nieh and others Olfactory eavesdropping by Trigona spinipes

active space of meliponine odour trails has not been s1 measured, although odours deposited at a feeder by 100 m Melipona panamica foragers can attract nest-mates over distances of 6 to 12 m (Nieh 1998). It is not known if interspecific meliponine eavesdropping actually occurs. Kerr et al. (1963) suggested that r2 fr2 Scaptotrigona xanthotricha may orient towards S. postica r1 fr1 odour marks. They trained colonies of S. postica and S. xanthotricha to separate feeders placed such that the odour trails from both colonies crossed. Two out of 122 S. postica foragers arrived at the S. xanthotricha feeder, fs whereas 28 out of 124 S. xanthotricha foragers arrived at s2 the S. postica feeder. These results are suggestive, but it is unclear how many foragers would have arrived at the Figure 1. Research site plan. Squares indicate colonies. Open circles indicate feeder sites. Lines connect colonies to feeder of the other species in the absence of odour marks. the feeder sites used with each colony. The star denotes the Sensitivity to interspecific odour marks may also work to global positioning system reference point: 21°26.387Ј S, the advantage of the excluded, allowing frequently 47°34.884Ј W. attacked species to detect aggressive species before serious attacks begin. Johnson (1974) noted that T. fulviventris 2. MATERIAL AND METHODS foragers appeared reluctant to land on food sites previously visited by the aggressive species, T. fuscipennis (a) Study site, feeders and training (Johnson & Hubbell 1974). However, it is not clear if any We used two colonies of T. spinipes (s1 and s2, ca. 8000 bees stingless bee can be attracted to or repulsed by the per colony) in trees and two colonies of M. rufiventris (r1 and food-marking odours deposited by an interspecific r2, 500–700 bees per colony) in hives at a ranch near Saõ Simaõ competitor. in the state of Saõ Paulo, Brazil, from August to September in We therefore chose to study the highly aggressive spec- 2002 and 2003. We studied the following pairs of colonies: 1 (s ,r); 1 (s ,r); 2 (s ,r); and 2 (s ,r) for a total of six trials ies T. spinipes, which defends and usurps food sources 1 1 1 2 2 1 2 2 from carpenter bees, Africanized honeybees and other per experiment. stingless bees (Cobert & Willmer 1980; Cortopassi-Laur- We trained 20 individually marked foragers from each M. rufiventris colony to feeders located 2 m east of each colony (f ) ino 1982; Cortopassi-Laurino & Ramalho 1988; Gallo et r (figure 1). We then trained 20 individually marked foragers from al. 1988; Sazima & Sazima 1989; Martinez & Bullock each T. spinipes colony to identical feeders located 4 m east of 1990; Ramalho et al. 1994; Silva et al. 1997), six species each M. rufiventris colony (f ). After one trial at f and one trial of passiform birds (Barbosa 1999) and several species of s s at f , we captured and removed all T. spinipes and M. rufiventris hummingbirds (Willmer & Corbet 1981; Gill et al. 1982). r foragers that had been trained to the feeders, and then trained Trigona spinipes foragers use cephalic glandular sections to a new group of foragers from each colony. We verified that all deposit odour trails and to odour-mark food sources (Kerr trained foragers came from the pair of colonies under study by 1972, 1973; Kerr et al. 1981). watching them enter their respective colony entrances. With Trigona spinipes is a foraging generalist (Barbola et al. each T. spinipes colony, we used aspirators (Nieh 1998) to cap- 2000) and inhabits diverse habitats, ranging from the cer- ture all foragers at the feeders and verifying for 3 h that no rado (neotropical savannah) to tropical forests throughout further foragers arrived at any feeder location before training for- South America (Schwarz 1948; Roubik 1989; Barros Hen- agers from the second colony. With M. rufiventris, we used wire riques 1997). Throughout its range, T. spinipes constructs mesh (applied in the evening after all bees had returned to their large external nests of mud, resin and wax, usually placed nest) to seal the entrance to the colony that was not under study. above the ground in large trees (Roubik 1989). Colonies Each feeder consisted of a small glass bottle (5 cm diameter, range in size from 5000 to over 100 000 workers 4.5 cm high, 65 ml) inverted over a grooved plastic base 6.7 cm (Michener 1974; Wille 1983; Almeida & Laroca 1988), in diameter (methods of Von Frisch 1967). To facilitate forager and thus T. spinipes form some of the largest stingless bee orientation, we placed a disc of yellow paper 6.7 cm in diameter colonies in the world (Roubik 1989). underneath all feeder bases. Each feeder contained unscented Our goal was to determine whether T. spinipes uses 2.5 M sucrose solution (Tautz & Sandeman 2003) and was olfactory eavesdropping to find and subsequently take placed on a grey plastic dish 20 cm in diameter supported by a over food sources from another species. We focus on the 1 m high tripod. aggressive interaction between T. spinipes and M. We define an experienced forager as any bee that has pre- rufiventris (Lepeletier 1835; Moure 1975), a moderately viously visited a feeder. All other foragers were newcomers. We aggressive stingless bee that may odour-mark food sources used paint pens to individually mark the thoraces of bees visiting and can pillage stingless bee nests for cerumen (wax), each feeder (Nieh et al. 2003a). We allowed a fixed number of propolis and honey (Kerr & Rocha 1988; Kerr 1994), but foragers to visit the feeder, and censused the number of marked generally does not attack or harass other species of sting- training-feeder foragers each 15 min, capturing or releasing less bees on floral resources (Rocha 1970; Souza 1978; marked foragers to maintain a constant number of recruiters to Breed & Page 1991). Both species occur in Amazonia, ensure a more controlled rate of recruitment. Marked and Brazil (Roubik 1983; Brown & Albrecht 2001), and we unmarked bees were captured in separate aspirators. At the end observed T. spinipes harassing and attacking M. rufiventris of each day, we released marked bees captured at the training on natural food sources at our field site. feeder and froze all unmarked T. spinipes foragers captured at

Proc. R. Soc. Lond. B (2004) Olfactory eavesdropping by Trigona spinipes J. C. Nieh and others 1635 both feeders. Thus unmarked bees could not return and be Table 1. Orientation of Melipona rufiventris foragers towards recounted at the feeder (Biesmeijer & de Vries 2001). To avoid odour marks (versus blanks) deposited by nest-mates on the depleting the much smaller M. rufiventris colony of foragers, we feeder. Each trial lasted 15 min and feeder positions were marked and released all captured unmarked M. rufiventris for- swapped every 5 min. agers at the end of each day and verified that they returned to number of foragers the colony under study. choosing (b) Odour mark collection trial colony experimental control two-tailed B.P. We collected putative M. rufiventris odour marks, T. spinipes odour marks and blanks that were not odour marked. To collect 1 1 13 0 0.0002 odour marks, we allowed 20 individually marked foragers to feed 2 1 15 0 Ͻ 0.000 for 30 min while standing on a ring of Whatman number 1 filter 3 1 21 0 0.00001 paper (5.5 cm inner diameter, 12 cm outer diameter) placed 4 1 34 1 0.00001 around the collection feeders (fs,fr1 and fr2). To obtain blanks, 5 2 17 0 Ͻ 0.0001 we placed a ring of paper for 30 min around an identical, but 6 2 22 1 0.0001 unvisited, feeder containing the same unscented sucrose solution 15 m west of the collection feeder. A monitor observed that no bees visited this feeder. We did not use filter paper that contacted M. rufiventris feeder from above to record how T. spinipes the sucrose solution or any odour-marked paper that contacted discovered and then excluded M. rufiventris during 9 min trials. foragers other than those of the subject colony. Experimenters In separate trials, we zoomed in to film 25% of the feeder and wore disposable gloves and used clean forceps and bags to handle recorded T. spinipes individual feeding durations (defined as forager the filter papers (methods in Nieh et al. 2003c). proboscis in sucrose solution). At the end of each trial, we captured and froze all T. spinipes foragers at the M. rufiventris feeders. (c) Testing the attractiveness of odour marks We tested the attractiveness of odour marks to T. spinipes new- (e) Video analysis comers and M. rufiventris experienced foragers in a 25 min To analyse and compare forager choice and orientation behav- paired-feeder assay. We conducted only one assay at a time, at a iour in detail, we filmed with a Canon XL-1 NTSC digital video single location. At the beginning of each trial, we placed odour- camera (30 frames sϪ1) positioned above the feeders (paired marked and non-odour-marked filter papers inside a clean, choice experiment) or feeder (competitive exclusion). We cap- sealed plastic bag and covered the training feeder with a plastic tured the data into an Apple PowerBook G4 computer using cylinder (taking care not to trap any foragers). After 5 min (filter Imovie v. 3.0.3 and used Videopoint v. 2.1 to measure the path paper transport time), we offered newcomers a choice between length, velocity and deceleration of T. spinipes foragers orienting two identical, empty, clean feeders. The last odour marks were towards the feeders (measured within 28 cm of the paired feeders). thus deposited a minimum of 5 min before presentation. Around In the competitive exclusion experiment, we counted the total each feeder, we placed one ring of filter paper. We placed feeders number of bees on the feeder, the number feeding and the number 8.5 cm to right and left of the original feeding site and captured fighting every 20 s (fighting as defined by Johnson (1974)). bees as soon as they landed. Owing to high recruitment rates (more than 100 newcomers (f ) Statistical analyses per hour), T. spinipes newcomers continued to arrive during this We use the ␹2-test to analyse the recruitment control trials. test phase even though all foragers, including the experienced In two-feeder experiments, we calculate probabilities from a foragers, were immediately captured before they could land, and two-tailed binomial distribution with p = q = 0.5 (binomial prob- no bees were therefore allowed to return to the nest to recruit. ability, B.P.). We use Mann–Whitney U-tests to analyse the We counted only individual choices made by newcomers in the flight orientation data and regression, and ANOVA for the com- absence of other bees. All T. spinipes newcomers were frozen or petitive exclusion data. All averages are reported as mean ±1 s.d. held inside a cage (left inside a closed room at the Fazenda) for the duration of the experiments. We released marked T. spinipes 3. RESULTS foragers at the end of each trial to recruit a new set of foragers. In feeder choice experiments with M. rufiventris, we tested the (a) Melipona rufiventris odour marking orientation of experienced M. rufiventris foragers (out of a pool Melipona rufiventris foragers deposited attractive odour of 120 individually marked foragers from each colony), releasing marks on the feeder (table 1). In all six trials, significantly them at the end of each trial. more M. rufiventris foragers chose the feeder with the filter Before testing the attraction of T. spinipes to feeders at the M. paper on which their nest-mates had formerly fed, over rufiventris sites (fr2 and fr1), we captured all M. rufiventris foragers the control feeder with no odour marks (p р 0.0002). and sealed the M. rufiventris colony. We also exchanged the positions of both feeders every 5 min to eliminate site bias, continu- (b) Trigona spinipes flight orientation ously monitored wind direction, and shifted the feeders so that Once we covered the T. spinipes training feeder, T. the axis connecting both feeders was always perpendicular to the spinipes foragers immediately began to search over an wind direction. increasingly wider area for an available food source. In this way, many T. spinipes foragers found and oriented towards (d) Competitive exclusion the M. rufiventris feeder sites. Figure 2a shows that T. When the odour-mark attraction experiment had been com- spinipes newcomers strongly preferred T. spinipes odours pleted, we removed the T. spinipes feeder (fs) and allowed M. marks to M. rufiventris odour marks at the T. spinipes rufiventris foragers to feed at either fr1 or fr2.Wefilmedthe feeder site (fs). However, T. spinipes newcomers searching

Proc. R. Soc. Lond. B (2004) 1636 J. C. Nieh and others Olfactory eavesdropping by Trigona spinipes

(a) foragers (marked bees) showed the same strong preferences, and during the 25 min of each attraction experiment, 77.5% chose T. spinipes odours over M. rufiventris

odours at fs, 75% chose M. rufiventris odours over T. spinipes odours at f and f , and 72.5% chose M. rufiventris sr r1 r2 odours over blanks at fr1 and fr2 (two-tailed B.P., p = 0.0001, n = 80 foragers per attraction experiment).

(d) M. rufiventris aversion Melipona rufiventris foragers showed a strong aversion to T. spinipes odour marks (figure 3c), avoiding these marks (b) at all tested time intervals and preferring M. rufiventris odour marks (99.6%, overall p 0.0001) or even a blank paper with no odour marks (100%, overall p 0.0001). sr (e) Competitive exclusion During the six exclusion trials (figure 4), the total number of T. spinipes foragers on the feeder and the number feeding significantly increased with time (ANOVA:

F1,26 у 23.3, p Ͻ 0.0001), and the number fighting sig-

nificantly decreased with time (ANOVA: F1,26 Figure 2. Flight paths of Trigona spinipes newcomers = 9.5, p = 0.005) as the number of M. rufiventris foragers orienting to odour marks: choice of T. spinipes (s)or decreased on the feeder. The total number of M. rufiv- Melipona rufiventris (r) odour marks. Filled circles denote the entris foragers on the feeder, the number feeding and the feeder bottles. Dashed circles indicate the filter papers. number fighting significantly decreased with time Preferential orientation (a)toT. spinipes odour marks at the у Ͻ T. spinipes feeder site (21 foragers) and (b)toM. rufiventris (ANOVA: F1,26 11.9, p 0.002). As T. spinipes for- odour marks at the M. rufiventris feeder site 1 (20 foragers). agers successfully took over the feeder, the amount of time Feeder positions were swapped every 5 min. Pooled results that T. spinipes foragers spent individually feeding also shown (six trials lasting 10 min each). Scale bar, 10 cm. increased (figure 4b; ANOVA: F1,17 = 11.8, p = 0.003). In all trials, T. spinipes foragers won and successfully excluded all M. rufiventris foragers (figure 4a). for food at the M. rufiventris feeder site (fr1) strongly preferred M. rufiventris odour marks to T. spinipes odours 4. DISCUSSION marks (figure 2b). Flight orientation behaviour was quite similar at both sites. There is no significant difference Trigona spinipes foragers can detect interspecific odour between the length (0.215 ± 0.149 m), average velocity marks deposited to advertise food sources to conspecifics. (0.201 ± 0.108 m sϪ1) or average deceleration (6.03 ± When searching for new feeding sites not already occupied Ϫ1 Ϫ1 3.23 m s s )ofT. spinipes orientation flight paths at fs by nest-mates, T. spinipes newcomers preferred M. or fr1 (Mann–Whitney U у 182, n1 = 20, n2 = 21, p rufiventris odour marks to nest-mate odour marks and dis- у 0.47). played the same flight orientation behaviour towards nest- mate-deposited and interspecific marks (figure 2). In the (c) T. spinipes attraction to odour marks competitive exclusion experiment, T. spinipes newcomers Figure 3 shows the differential responses of T. spinipes immediately began fighting if M. rufiventris foragers newcomers in greater detail. At their own feeder site (fs), occupied the feeder, behaving as extirpators and dis- significantly more T. spinipes newcomers preferred the playing all four levels of aggression categorized by Johnson odour marks of their nest-mates to those of M. rufiventris, (1974): threats to intense grappling followed by decapi- even after 20 min had passed (overall p 0.0001). The tations. In all six trials, this strategy was successful and T. proportion of newcomers choosing T. spinipes odour spinipes quickly won control after driving away or killing marks steadily decreased with time as the odour marks all M. rufiventris foragers (figure 4). evaporated (figure 3b). The strong aversion shown by M. rufiventris foragers

At the M. rufiventris feeder sites (fr1 and fr2), T. spinipes towards T. spinipes odour marks (figure 3c) supports the newcomers preferred M. rufiventris odour marks to those of hypothesis that some species may detect interspecific their nest-mates within the first 15 min, with no preference odour marks to avoid confrontations rather than to exploit exhibited thereafter (figure 3b,overall p 0.0001). When discoveries of other bees (Johnson 1974; Hubbell & John- given a choice of M. rufiventris odour marks and blanks (no son 1978). This avoidance response is in sharp contrast to odour marks), T. spinipes newcomers preferred M. rufiventris the preferences of M. rufiventris foragers choosing between odour marks within the first 20 min (overall p 0.0001). M. rufiventris odour marks and no odour marks. In this In both cases, attraction to the odour marks steadily case, almost no foragers (1.6%) chose the blank (table 1). decreased with time as the odour marks evaporated. Moreover, the aversion was evident the first time that each Thus T. spinipes newcomers could clearly distinguish M. rufiventris colony experienced T. spinipes odour marks between T. spinipes odour marks and M. rufiventris odour (figure 3c; 100% avoidance of T. spinipes odours versus marks, but strongly preferred M. rufiventris odour marks blanks in all trials). Whether this aversion is a result of prior when searching away from fs. Experienced T. spinipes experiences with T. spinipes foragers at other sites, is specific

Proc. R. Soc. Lond. B (2004) Olfactory eavesdropping by Trigona spinipes J. C. Nieh and others 1637

r Figure 3. Testing attraction to interspecific odour marks (each plot shows pooled data from six trials). Bars show the (a) 100 number of foragers choosing a particular odour type (black bars, Trigona spinipes odour; hatched bars, Melipona 80 300 *** rufiventris odour) or control (white bars). The line plot with 60 s.d. bars shows the average percentage of foragers choosing the experimental feeder. Two-tailed B.P.: ∗∗∗ p 0.0001, 200 *** 40 ∗∗ p Ͻ 0.001, ∗ p = 0.05. Attraction of T. spinipes newcomers 20 (a) to odour marks at the T. spinipes feeder site and (b)to *** odour marks at the M. rufiventris feeder sites: (i) shows 100 *** * 0 choice between T. spinipes and M. rufiventris odour marks;

number of foragers (ii) shows choice between M. rufiventris and no odour marks. (c) Aversion of M. rufiventris foragers to T. spinipes odour 0 marks (zero values shown with a ‘0’): (i) shows choice percentage choosing experimental feede between T. spinipes and M. rufiventris odour marks; (ii) (b)(i) shows choice between T. spinipes and no odour marks. 100

*** 80 40 60 to T. spinipes odour marks or is a general response to foreign ** odour marks remains to be determined. ** 40

20 20 (a) Newcomer identity 0 The colony identity of T. spinipes newcomers that were frozen was not verified, but it is unlikely that these foragers 0 came from non-subject colonies because we did not 100 observe conspecific fighting or stereotyped fleeing behav- (ii) iour, as occurs when T. spinipes foragers from different 120 80 colonies meet on a food source (P. Nogueira-Neto, per- number of foragers *** 60 sonal communication). Moreover, T. spinipes experienced 80 foragers that were directly verified as coming from the col- 40 *** onies under study (the marked foragers), showed the same

20 percentage choosing experimental feeder pattern of odour choice exhibited by T. spinipes new- 40 *** ** comers at the T. spinipes and M. rufiventris feeder sites. 0 0 (b) Attraction by other means? 0 5 10 15 20 It is important to consider alternative explanations for our results. Could T. spinipes newcomers orienting minutes towards other sources of information, aside from M. rufiv- (c)(i) entris odours, account for our data? We counted only T. 100 spinipes newcomers that had never previously experienced 120 *** 80 a feeder, counted only bees that arrived individually and made a choice in the absence of other bees, and provided 60 identical feeders that we rotated every 5 min during the 80 40 test phases. Thus visual orientation to other bees (local enhancement; see Slaa et al. (2003)) and potential direc- 40 *** 20 tional biases do not account for our results. Exchanging *** *** *** 0 the positions of both feeders every 5 min also eliminated the possibility of biases as a result of minute differences 0 000 0 (less than 17 cm) in locale odours. 100 Because we used unscented sucrose solution and pro- (ii) vided empty feeder bottles during the test phases, T. 50 *** 80 number of foragers spinipes newcomers could not have oriented to sucrose sol- 40 60 ution odours. Trigona spinipes newcomers may have been attracted to the M. rufiventris feeder sites by nest-mate 30 *** *** 40

*** percentage choosing experimental feeder odours when offered a choice between T. spinipes odours 20 20 and M. rufiventris odours. However, T. spinipes newcomers *** were equally attracted to these sites when presented with 10 0 a choice between M. rufiventris odours and blanks (figure 0 0000 0 3b). In addition, significantly more T. spinipes newcomers 0 5 10 15 20 preferred to land on M. rufiventris odour marks in the first minutes 15 min of all trials and never showed a preference for T. spinipes odours at any time interval during any trial at the

M. rufiventris feeder sites (fr1 and fr2).

Proc. R. Soc. Lond. B (2004) 1638 J. C. Nieh and others Olfactory eavesdropping by Trigona spinipes

(a)(i) 60 It is possible that one aspect of locale odour, the odour of feeder assistants, drew some newcomers to fr1 and fr2. However, feeder assistants positioned themselves equidis- tant to both feeders, and thus locale odour does not 40 account for the strong preferences of T. spinipes newcomers for M. rufiventris odour marks at fr1 and fr2. Finally, both M. rufiventris feeder sites were closer to the T. spinipes colonies than the T. spinipes feeder (figure 20 1). Thus T. spinipes scouts searching for new food sources may have arrived at fr1 and fr2 without necessarily orienting to the M. rufiventris odours. Nonetheless, once they arrived, T. spinipes foragers displayed strong, consistent 0 and highly significant preferences for M. rufiventris odours (ii) 60 at the M. rufiventris feeder sites.

(c) Inability to distinguish odours? Could T. spinipes be unable to distinguish between con- 40 specific and interspecific odour marks? Trigona spinipes foragers odour mark with cephalic gland secretions (Kerr 1972, 1973; Kerr et al. 1981). The source of M. rufiventris odour marks is unknown, but other species of Melipona 20 odour mark with tarsal gland secretions and anal droplets (Nieh et al. 2003c; Hrncir et al. 2004a). At all locations, number of individuals each species was clearly able to distinguish between its odour marks and those of the different species (figure 3). 0 Moreover, postulating no differences in odour-mark com- 10 position, but differences in quantity, does not consistently account for the preferences observed. For example, if both (iii) species produced odour marks with the same chemical composition, but M. rufiventris produced them in greater quantity, T. spinipes foragers should have preferred M. rufiventris odour marks at all feeder sites. This was not the 5 case (figure 3a,b). Conversely, if T. spinipes odour marks are chemically identical to M. rufiventris odour marks, but are produced in greater quantity, T. spinipes foragers should have preferred their own odour marks at all sites. This also did not occur (figure 3a,b).

0 (d) Evolutionary implications (b) In bees, olfactory signalling provides an opportunity for 100 eavesdropping: a reasonable strategy because new feeding sites should be sought once old sites become exhausted or have sufficient labour allocated (Waddington & Holden 1979; Seeley 1995). Exploiting the discoveries of other 10 species (Johnson 1974) by orienting to their communication signals could provide a ready means to find rich new food sources. Floral resources are generally more scattered and yield individually poorer rewards than our 1 ad libitum feeders; however, dense inflorescences provided by large blooming tropical trees (such as Cassia bicapsularis log(feeding durations (s)) R2 = 0.41 at our field site) can provide a rich food source that takes 0.1 time to fully exploit and may thus be rewarding to eaves- droppers. In addition, several meliponine species, includ- 0 200 400 600 ing T. spinipes, can discover and raid weaker bee colonies, time from start of exclusion (s) extremely rich food sources that are highly sought after and sometimes contested with other raiders (Sakagami et Figure 4. (a) Trigona spinipes competitively excludes al. 1993; Nogueira-Neto 1997). Melipona rufiventris from the M. rufiventris feeder (pooled If olfactory eavesdropping exerted a selective pressure, data from six trials). (i) On feeder; (ii) feeding; and (iii) fighting. Filled circles indicate T. spinipes foragers. Open less conspicuous odour-marking strategies should have circles indicate M. rufiventris foragers. (b) Relationship evolved. Stingless bees use a range of olfactory recruitment between T. spinipes individual feeding durations and the time strategies: complete odour trails, short odour trails and from the start of exclusion (pooled data from six close-up point-source marking of the food source alone (Lindauer & video trials, log regression line shown). Kerr 1958; Nieh & Roubik 1995; Nieh et al. 2003c).

Proc. R. Soc. Lond. B (2004) Olfactory eavesdropping by Trigona spinipes J. C. Nieh and others 1639

Point-source marking is still susceptible to eavesdropping Cobert, S. A. & Willmer, P. G. 1980 Pollination of the yellow (figures 2 and 3), but may be less conspicuous than com- passion fruit: nectar, pollen and carpenter bees. J. Agric. Sci. plete odour trails. It is thus relevant to consider the adaptive 95, 655–666. value of these strategies with regard to eavesdropping. Cortopassi-Laurino, M. 1982 Divisaõ de recursos tro´ficos In particular, some meliponine species may use func- entre abelhas sociais, principalmente em Apis mellifera Linn. e Trigona (Trigona) spinipes Fabricius (Apidae, tionally referential communication (Nieh 2004). Although Hymenoptera). PhD thesis, Universidade de Saõ Paulo, Bra- the existence of such communication has not been con- zil, p. 180. clusively demonstrated in stingless bees (Hrncir et al. Cortopassi-Laurino, M. & Ramalho, M. 1988 Pollen harvest 2004b), studies have found correlations between distance by Africanized Apis mellifera and Trigona spinipes in Saõ and even the height of food sources (Nieh & Roubik 1998) Paulo: botanical and ecological views. Apidologie 19, 1–24. and the temporal structure of recruitment sounds pro- Dyer, F. C. 2002 The biology of the dance language. A. Rev. duced inside the nest in several species (Esch et al. 1965; Entomol. 47, 917–949. Esch 1967; Aguilar & Bricenõ 2002; Nieh et al. 2003b). Eltz, T., Bru¨hl, C. A., Van der Kaars, S., Chey, V. K. & Interestingly, these same species, when they have been Linsenmair, K. E. 2001 Pollen foraging and resource par- examined, all appear to use point-source odour marking titioning of stingless bees in relation to flowering dynamics (Nieh 1998; Hrncir et al. 2000, 2004a; Esch et al. 2001), in a southeast Asian tropical rainforest. Insectes Sociaux 48, and the only other highly social bees, honeybees, also use 273–279. Eltz, T., Bru¨hl, C. A., Van der Kaars, S. & Linsenmair, K. E. point-source odour marking in conjunction with the refer- 2002 Determinants of stingless bee nest density in lowland ential waggle dance (Esch et al. 2001; Dyer 2002). dipterocarp forests of Sabah, Malaysia. Oecologia 131, 27–34. Honeybees evolved in tropical habitats (Michener 2000) Esch, H. 1967 Die Bedeutung der Lauterzeugung fu¨r die Ver- in which they probably competed and still compete with sta¨ndigung der stachellosen Bienen. Zeitschrift fu¨r Vergleich- aggressive stingless bees (Nagamitsu & Inoue 1997). Thus ende Physiologie 56, 408–411. it remains unclear what forces have driven the evolution Esch, H., Esch, I. & Kerr, W. E. 1965 Sound: an element com- of functionally referential communication system in bees, mon to communication of stingless bees and to dances of but eavesdropping could have contributed to the evolution the honey bee. Science 149, 320–321. of location information encoded and transmitted at the Esch, H. E., Zhang, S., Srinivasan, M. V. & Tautz, J. 2001 well-defended nest (Nieh 1999). Honeybee dances communicate distances measured by optic flow. Nature 411, 581–583. The authors thank Paulo Nogueira-Neto, Juliana Rangel and Gallo, D., Nakano, O., Silveira Neto, S., Carvalho, R. P. L., de John Kim for their assistance, and Mark Hauber, David Holway Batista, G. C., Berti Filho, E., Parra, J. R. P., Zucchi, R. A., and anonymous reviewers for their comments on the manu- Alves, S. B. & Vendramim, J. D. 1988 Manual de entomologia script. Financial support came from NSF grant 0316697, UC agrıćola.Saõ Paulo, Brazil: Agronoˆmica Ceres. Academic Senate grant RC077C-NIEH, the Heiligenberg Chair Gill, F. B., Mack, A. L. & Russell, T. R. 1982 Competition Endowment at UCSD, and FAPESP 02/00582-5 of Brazil. between hermit hummingbirds Phaethorninae and insects for nectar in a Costa Rican rainforest. Ibis 124, 44–49. Hrncir, M., Jarau, S., Zucchi, R. & Barth, F. G. 2000 Recruit- REFERENCES ment behavior in stingless bees, Melipona scutellaris and M. quadrifasciata. II. Possible mechanisms of communication. Aguilar, I. & Bricenõ, D. 2002 Sounds in M. costaricensis Apidologie 31, 93–113. (Apidae: Meliponini): effect of sugar concentration and nec- Hrncir, M., Jarau, S., Zucchi, R. & Barth, F. G. 2004a On the tar source distance. Apidologie 33, 375–388. origin and properties of scent marks deposited at the food Almeida, M. C. & Laroca, S. 1988 Trigona spinipes (Apidae, source by a stingless bee, Melipona seminigra. Apidologie 35, Meliponinae): taxonomy, bionomy, and trophic relationships 3–13. in restricted areas. Acta Biolo´gica Paranaense 17, 67–108. Hrncir, M., Jarau, S., Zucchi, R. & Barth, F. G. 2004b Thorax Barbola, I. F., Laroca, S. & Almeida, M. C. 2000 Utilizaçaõ de recursos florais por abelhas silvestres (Hymenoptera, vibrations of a stingless bee (Melipona seminigra). I. No influ- Apoidea) da Floresta Estadual Passa Dois (Lapa, Parana´, ence of visual flow. J. Comp. Physiol. A. (In the press.) Brasil). Revista Brasileira de Entomologia, Saõ Paulo 44, 9–19. Hubbell, S. P. & Johnson, L. K. 1978 Comparative foraging Barbosa, A. A. A. 1999 Hortia brasiliana Vand. (Rutaceae): behavior of six stingless bee species exploiting a standardized pollination by Passeriformes in cerrado, southeastern Brazil. resource. Ecology 59, 1123–1136. Revista Brasileira de Botaˆnica 22, 99–105. Johnson, L. K. 1974 The role of agonistic behavior in the for- Barros Henriques, R. P. 1997 Nest density of Trigona spinipes aging strategies of Trigona bees. PhD thesis, University of (Hymenoptera: Apidae) in cerrado vegetation of central Bra- California Berkeley, Berkeley, CA, USA. zil. Revista de Biologia 44/45, 700–701. Johnson, L. K. & Hubbell, S. P. 1974 Aggression and compe- Biesmeijer, J. C. & de Vries, H. 2001 Exploration and exploi- tition among stingless bees: field studies. Ecology 55, 120–127. tation of food sources by social insect colonies: a revision of Kerr, W. E. 1960 Evolution of communication in bees and its the scout–recruit concept. Behav. Ecol. Sociobiol. 49, 89–99. role in speciation. Evolution 14, 326–327. Blumstein, D. M. 1999 The evolution of functionally referen- Kerr, W. E. 1972 Orientaçaõ pelo sol em Trigona spinipes. tial alarm communication: multiple adaptations; mutiple Cieˆncia e Cultura (Suppl.), 341–342. constraints. Evol. Commun. 3, 135–147. Kerr, W. E. 1973 Sun compass orientation in the stingless Breed, M. D. & Page, R. E. J. 1991 Intra- and interspecific bees, Trigona (Trigona) spinipes (Fabricius, 1793) (Apidae). nestmate recognition in Melipona workers (Hymenoptera: Anais da Academia Brasileira de Cieˆncias 45, 301–308. Apidae). J. Insect Behav. 4, 463–469. Kerr, W. E. 1994 Communication among Melipona workers Brown, J. C. & Albrecht, C. 2001 The effect of tropical defor- (Hymenoptera: Apidae). J. Insect Behav. 7, 123–128. estation on stingless bees of the genus Melipona (Insecta: Kerr, W. E. & Rocha, R. 1988 ComunicaçaõemMelipona Hymenoptera: Apidae: Meliponini) in central Rondonia, rufiventris e Melipona compressipes. Cieˆncia e Cultura 40, Brazil. J. Biogeogr. 28, 623–634. 1200–1203.

Proc. R. Soc. Lond. B (2004) 1640 J. C. Nieh and others Olfactory eavesdropping by Trigona spinipes

Kerr, W. E., Ferreira, A. & Simo˜es de Mattos, N. 1963 Com- Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabels- munication among stingless bees: additional data teen, T. 2002 Do great tits assess rivals by combining direct (Hymenoptera: Apidae). J. N. Y. Entomol. Soc. 71, 80–90. experience with information gathered by eavesdropping? Kerr, W. E., Blum, M. & Fales, H. M. 1981 Communication Proc. R. Soc. Lond. B 269, 1925–1929. (DOI 10.1098/ of food sources between workers of Trigona (Trigona) rspb.2002.2112) spinipes. Revista Brasileira de Biologia 41, 619–623. Ramalho, M., Giannini, C., Malagodi-Braga, K. S. & Lepeletier, S.-F. d. 1835 Histoire naturelle d’insectes-hymeńopte´res. Imperatriz-Fonseca, V. L. 1994 Pollen harvest by stingless Paris: Roret. bee foragers (Hymenoptera, Apidae, Meliponinae). Grana Lindauer, M. & Kerr, W. E. 1958 Die gegenseitige Versta¨ndi- 33, 239–244. gung bei den stachellosen Bienen. Zeitschrift fu¨r Vergleichende Rocha, E. S. M. O. 1970 Raiding in Melipona rufiventris flavoli- Physiologie 41, 405–434. neata. In Proc. 22nd Ann. Meeting of the Sociedade Brasileira Liow, L. H., Sodha, N. S. & Elmqvist, T. 2001 Bee diversity Para o Progresso da Ciencia (ed. S. Ceccato), p. 292. Brazil: along a disturbance gradient in tropical lowland forests of Sociedade Brasileira Para o Progresso da Ciencia. southeast Asia. J. Appl. Ecol. 38, 180–192. Roubik, D. W. 1982 Seasonality in colony food storage, brood McGregor, P. K. 1993 Signalling in territorial systems: a con- production, and adult survivorship: studies of Melipona in text for individual identification, ranging and eavesdropping. tropical forest (Hymenoptera: Apidae). J. Kansas Entomol. Phil. Trans. R. Soc. Lond. B 340, 237–244. Soc. 55, 789–800. Marler, P., Evans, C. S. & Hauser, M. D. 1992 Animal signals: Roubik, D. W. 1983 Thermodynamics in nests of two Melipona motivation, referential, or both? In Nonverbal vocal communi- species in Brazil. Acta Amazonica 13, 453–471. cation: comparative and developmental approaches (ed. M. Roubik, D. W. 1989 Ecology and natural history of tropical bees. Papoucek), pp. 66–86. Cambridge University Press. New York: Cambridge University Press. Martinez, D. E. & Bullock, S. H. 1990 Floral parasitism by Roubik, D. W. & Aluja, M. 1983 Flight ranges of Melipona social bees (Meliponinae, Apidae) on the chiropterophilous and Trigona in tropical forest. J. Kansas Entomol. Soc. 56, tree Crescentia alata (Bignoniaceae). Boletin de la Sociedad 217–222. Botanica de Mexico 50, 69–76. Sakagami, S. F., Roubik, D. W. & Zucchi, R. 1993 Ethology Michener, C. D. 1974 The social behavior of the bees. of the robber stingless bee, Lestrimelitta limao (Hymenoptera: Cambridge, MA: Harvard University Press. Apidae). Sociobiology 21, 237–277. Michener, C. D. 2000 The bees of the world. Baltimore, MD: Sazima,I.&Sazima,M.1989Carpenterbeesandstingless Johns Hopkins University Press. honeybees (Hymenoptera, Apoidea): visiting, interactions and Moure, J. C. 1975 Notas sobre as espećies de Melipona consequences for the pollination of the passionflower descritas por Lepeletier em 1836 (Hymenoptera-Apidae). (Passifloraceae). Revista Brasileira de Entomologia 33, 109–118. Revista Brasileira de Biologia 35, 615–623. Schwarz, H. F. 1948 Stingless bees (Meliponinae) of the West- Nagamitsu, T. & Inoue, T. 1997 Aggressive foraging of social ern Hemisphere. Bull. Am. Mus. Nat. Hist. 90, 1–546. bees as a mechanism of floral resource partitioning in an Seeley, T. D. 1995 The wisdom of the hive: the social physiology of Asian tropical rainforest. Oecologia 110, 432–439. honeybee colonies. Cambridge, MA: Harvard University Press. Nieh, J. C. 1998 The role of a scent beacon in the communi- Silva, M. M., Buckner, C. H., Picanço, M. & Cruz, C. D. cation of food location in the stingless bee, Melipona panamica. 1997 Influeˆncia de Trigona spinipes Fabr. (Hymenoptera: Behav. Ecol. Sociobiol. 43, 47–58. Apidae) na Polinizaçaõ do Maracujazeiro Amarelo. Anais da Nieh, J. C. 1999 Stingless-bee communication. Am. Sci. 87, Sociedade Entomologica do Brasil 26, 217–221. 428–435. Slaa, E. J., Wassenberg, J. & Biesmeijer, J. C. 2003 The use Nieh, J. C. 2004 Recruitment communication in stingless bees of field-based social information in eusocial foragers: local (Hymenoptera, Apidae, Meliponini). Apidologie 35,159–182. enhancement among nestmates and heterospecifics in sting- Nieh, J. C. & Roubik, D. W. 1995 A stingless bee (Melipona less bees. Ecol. Entomol. 28, 369–379. panamica) indicates food location without using a scent trail. Souza, S. C. F. 1978 Notes on pillage behavior of Melipona Behav. Ecol. Sociobiol. 37, 63–70. rufiventris flavolineata (Hymenoptera, Apoidea). Revista Nieh, J. C. & Roubik, D. W. 1998 Potential mechanisms for Brasileira de Entomologia 22, 95–98. the communication of height and distance by a stingless bee, Stowe, M. K., Turlings, T. C. J., Loughrin, J. H., Lewis, Melipona panamica. Behav. Ecol. Sociobiol. 43, 387–399. W. J. & Tumlinson, J. H. 1995 The chemistry of eavesdrop- Nieh, J.C., Contrera, F.A.L. & Nogueira-Neto, P. 2003a ping, alarm, and deceit. Proc. Natl Acad. Sci. USA 92,23–28. Pulsed mass-recruitment by a stingless bee, Trigona hyalinata. Tautz, J. & Sandeman, D. C. 2003 Recruitment of honeybees Proc. R. Soc. Lond. B 270, 2191–2196. (DOI 10.1098/ to non-scented food sources. J. Comp. Physiol. A 189, rspb.2003.2486.) 293–300. Nieh, J. C., Contrera, F. A. L., Rangel, J. & Imperatriz- Van Nieuwstadt, M. G. L. & Ruano, I. C. E. 1996 Relation Fonseca, V. L. 2003b Effect of food location and quality on between size and foraging range in stingless bees (Apidae, recruitment sounds and success in two stingless bees, Meliponinae). Apidologie 27, 219–228. Melipona mandacaia and Melipona bicolor. Behav. Ecol. Von Frisch, K. 1967 The dance language and orientation of bees. Sociobiol. 55, 87–94. Cambridge, MA: Belknap Press. Nieh, J. C., Ramı´rez, S. & Nogueira-Neto, P. 2003c Multi- Waddington, K. D. & Holden, L. R. 1979 Optimal foraging: source odor-marking of food by a stingless bee, Melipona on flower selection by bees. Am. Nat. 114, 179–196. mandacaia. Behav. Ecol. Sociobiol. 54, 578–586. Whitfield, J. B. 2002 Nosy neighbours. Nature 419, 242–243. Nogueira-Neto, P. 1997 Vida e criaçaõ de abelhas indı´genas sem Wille, A. 1983 Biology of the stingless bees. A. Rev. Entomol. ferraõ.Saõ Paulo, Brazil: Editoria Nogueirapis. 28, 41–64. Oliveira, R. F., McGregor, P. K. & Latruffe, C. 1998 Know Willmer, P. G. & Corbet, S. A. 1981 Temporal and microcli- thine enemy: fighting fish gather information from observing mactic partitioning of the floral resources of Justica aurea conspecific interactions. Proc. R. Soc. Lond. B 265, 1045– amongst a concourse of pollen vectors and nectar robbers. 1049. (DOI 10.1098/rspb.1998.0397.) Oecologia 51, 67–78. Peake, T. M. 2004 Eavesdropping in communication networks. In Animal communication networks (ed. P. K. As this paper exceeds the maximum length normally permitted, the McGregor). Cambridge University Press. (In the press.) authors have agreed to contribute to production costs.

Proc. R. Soc. Lond. B (2004)